Article
Effects of Co-Worker and Supervisor Support on Job Stress and Presenteeism in an Aging Workforce: A Structural Equation Modelling Approach Tianan Yang 1 , Yu-Ming Shen 2 , Mingjing Zhu 3 , Yuanling Liu 4 , Jianwei Deng 5 , Qian Chen 6 and Lai-Chu See 7,8, * Received: 25 October 2015; Accepted: 18 December 2015; Published: 23 December 2015 Academic Editor: Cary Cooper 1 2
3 4 5 6 7 8
*
Department of Organization and Human Resource Management, School of Management and Economics, Beijing Institute of Technology, Beijing 100081, China; [email protected] Institute of Medical Informatics, Biometry and Epidemiology (IBE), Pettenkofer School of Public Health, Faculty of Medicine Ludwig-Maximilians University Munich, Munich 81377, Germany; [email protected] Institute of Psychology, Chinese Academy of Sciences, Beijing 100101, China; [email protected] Human Resources Department, Guangdong Women and Children Hospital, Guangzhou 511442, China; [email protected] Department of Public Administration, School of Management and Economics, Beijing Institute of Technology, Beijing 100081, China; [email protected] Medical Affair Department, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences, Beijing 100032, China; [email protected] Department of Public Health, Medical College, Chang Gung University, Taoyuan 33302, Taiwan Biostatistics Core Laboratory, Molecular Medicine Research Center, Chang Gung University, Taoyuan 33302, Taiwan Correspondence: [email protected]; Tel.: +886-3-211-8800 (ext. 5119); Fax: +886-3-211-8363
Abstract: We examined the effects of co-worker and supervisor support on job stress and presenteeism in an aging workforce. Structural equation modelling was used to evaluate data from the 2010 wave of the Health and Retirement Survey in the United States (n = 1649). The level of presenteeism was low and the level of job stress was moderate among aging US workers. SEM revealed that co-worker support and supervisor support were strongly correlated (β = 0.67; p < 0.001). Job stress had a significant direct positive effect on presenteeism (β = 0.30; p < 0.001). Co-worker support had a significant direct negative effect on job stress (β = ´0.10; p < 0.001) and presenteeism (β = ´0.11; p < 0.001). Supervisor support had a significant direct negative effect on job stress (β = ´0.40; p < 0.001) but not presenteeism. The findings suggest that presenteeism is reduced by increased respect and concern for employee stress at the workplace, by necessary support at work from colleagues and employers, and by the presence of comfortable interpersonal relationships among colleagues and between employers and employees. Keywords: co-worker support; equation modelling
supervisor support;
job stress;
presenteeism;
structural
1. Introduction Presenteeism can be viewed from two distinct perspectives. It is, of course, the literal antonym of absenteeism, and UK and European researchers in management, epidemiology, and occupational health [1–3] have investigated this positive aspect of the term, i.e., when employees remain at work, even when they are sick, or overstate their attendance, because of job insecurity due to downsizing
Int. J. Environ. Res. Public Health 2016, 13, 72; doi:10.3390/ijerph13010072
www.mdpi.com/journal/ijerph
Int. J. Environ. Res. Public Health 2016, 13, 72
2 of 15
and restructuring forces. However, presenteeism can be regarded as an indicator of lost work productivity, and US medical researchers and consultants [4] are concerned about the adverse effects that sickness and specific medical conditions might have on work productivity in organizations. In this article, we discuss presenteeism from the perspective of lost work productivity, because such analysis can address gray gaps between the absence of productivity and full productivity [5] and between healthy and unhealthy persons [6]. Moreover, previous studies suggest that further research is needed in order to better understand the potential differences between absenteeism and presenteeism [7–11]. Presenteeism was initially defined in the field of occupational health as the act of attending work while unable to perform effectively due to health problems [12,13]. The increasing attentions has been attracted is because presenteeism is responsible for 3 times and 1.8 times the financial burden of medical illness and absenteeism, respectively [14]. However, with the development of definition of presenteeism, more recently the definition has been extended to include other conditions and events that limit productivity, as suggested by Johns [15]. Advances in public health, medicine, science, and technology have greatly improved the health and life expectancy in developed and developing countries [16,17]. However, although retirement age is now older [18,19], physical and cognitive capabilities nevertheless deteriorate with age. This could lead to productivity loss and increased employer concerns regarding organizational competitiveness. Unfortunately, the extent of presenteeism among aging workers is not well understood. Employees who have demanding jobs, low decision latitude, job strain, and/or low social support tend to have more sickness absences [20] and, consequently, severe presenteeism [21–23]. Job stress occurs when an individual ability cannot meet job demands [24,25]. The job demand–resources model (JD-R) holds that when job demands are high and there are few job resources, employees may suffer more job stressors, which results in high job stress and other consequences. Because job demands must be satisfied if a worker is to remain employed, employees pretend to work hard at the workplace and forgo absences, even while they are sick or not fully productive. The greater the job requirements, the more efforts employees make in meeting them and the greater the probability that they will work while sick, to ensure full-time presence [26]. Fortunately, strong support from co-workers and supervisors improves work environments by relieving employee stress [27,28], which enhances job satisfaction and performance [29] and subsequently reduces presenteeism in enterprises and organizations [30,31]. Supervisors are in positions that can address employee complaints and help employees obtain necessary resources [32]. Co-workers can successfully finish work tasks and reduce stress and presenteeism [33]. In agreement with the buffering model of social support, Cummins [34] reported that employees that had good relationships with supervisors and co-workers are usually successful and productive at work, even when job stress is severe. Although co-worker support and supervisor support are both important in reducing job stress [35], most studies have investigated these two support mechanisms separately [36]. In addition, the relationship between co-worker and supervisor support has rarely been studied. Moreover, the level of co-worker and supervisor support that employees in an aging workforce receive is unclear. Most previous studies [27,34] used linear regression to investigate relationships between co-worker support, supervisor support, job stress, and presenteeism; however, such analysis cannot account for the complicated relationships among these variables. In this study, we used structural equation modelling (SEM) to examine the complicated effects of co-worker and supervisor support on job stress and presenteeism in an aging workforce. We hypothesized that there would be a direct buffer effect of supervisor support on job stress and presenteeism, a direct buffer effect of co-worker support on job stress and presenteeism, and a direct positive effect of job stress on presenteeism (Figure 1).
Int. J. Environ. Res. Public Health 2016, 13, 72 Int. J. Environ. Res. Public Health 2016, 13, 0072
3 of 15
Figure 1. 1. Initial Initial model model of of how how co-worker co-worker and and supervisor supervisor support support affect affect job job stress stress and and presenteeism. presenteeism. Figure
2. Method 2. Method 2.1. Data Data Source Source 2.1. We conducted conducted aa secondary secondary analysis analysis of of data data from from the the 2010 2010 wave wave of of the the Health Health and and Retirement Retirement We Survey (HRS) (HRS) in in the the United United States. States. The measures the the health, health, retirement, retirement, psychosocial psychosocial factors, factors, Survey The HRS HRS measures and productivity of the labour market and is funded by the National Institute of Aging and the Social and productivity of the labour market and is funded by the National Institute of Aging and the Social Security Administration of the United States. The HRS was initiated in 1992 and recruited persons Security Administration of the United States. The HRS was initiated in 1992 and recruited persons older than than 50 50 years years for for participation participation in in biennial biennial surveys surveys based based on on aa multistage multistage area area probability probability older sample. To avoid the aging problem and a decrease in the number of participants over time, new new sample. To avoid the aging problem and a decrease in the number of participants over time, samples were were added added every every66 years years [37,38]. [37,38]. In In 2006, 2006, the the HRS HRS added added aa participant participant lifestyle lifestyle questionnaire questionnaire samples (PLQ) to their core biennial survey, for a random 50% of the core panel participants. The The PLQPLQ was (PLQ) to their core biennial survey, for a random 50% of the core panel participants. developed by the HRS Psychosocial Working Group and includes a perceived ability to work scale was developed by the HRS Psychosocial Working Group and includes a perceived ability to work (PAWS), job stress scale, co-worker supportsupport scale, and supervisor support scale [39]. Detailed information scale (PAWS), job stress scale, co-worker scale, and supervisor support scale [39]. Detailed on the study population and survey methodology has been published elsewhere [40]. information on the study population and survey methodology has been published elsewhere [40]. In the the 2010 2010 wave wave of of the the HRS, HRS, PLQ PLQ data data were were available available for for 8080 8080 participants, participants, because because only only In respondents in in 2010 2010 who who had had completed completed the the face-to-face face-to-face PLQ PLQ interview interview in in 2006 2006 again again rotated rotated to to this this respondents mode of data collection [37,38]. Among these 8080 participants, 2730 (33.8%) were still employed and mode of data collection [37,38]. Among these 8080 participants, 2730 (33.8%) were still employed and older than than 50 50 years. years. Among Among those those still still employed, employed, 1649 1649 (60.4%) (60.4%) answered answered at at least least one one question question on on the the older “PLQ 2010”, and the percentage of those with missing data was less than 2.6%. Because presenteeism “PLQ 2010”, and the percentage of those with missing data was less than 2.6%. Because presenteeism is only only observed observed among among employed employed persons, persons, data data from from these these 1649 1649 participants participants were were analyzed analyzed in in this this is study (Table (Table1). 1). study Table 1.1. Demographic Demographic characteristics characteristics of of the the total total sample sample of of aging aging workers workers and and the the subset subset with with data data Table from the participant lifestyle questionnaire (this study) in the 2010 wave of the Health and Retirement from the participant lifestyle questionnaire (this study) in the 2010 wave of the Health and Retirement Survey,USA. USA. Survey, Characteristics Characteristics Sex Sex Male Male Female Female Age (years) 50–59 Age (years) 50–59 60–69 60–69 70–79 70–79 ≥80 ě80 Education NoEducation degree No degree Grade 1–9 Grade 1–9 High school diploma Two-year college degree Four-year college degree Master’s degree
Total Sample (n = 2730) Total Sample (n = 2730)
This Study (n = 1649) This Study (n = 1649)
1237 14931237 1493
(45.3%) (45.3%) (54.7%) (54.7%)
750 750 899 899
(45.5%) (45.5%) (54.5%) (54.5%)
1520 8641520 309864 37 309 37
(55.7%) (55.7%) (31.6%) (31.6%) (11.3%) (11.3%) (1.4%) (1.4%)
990 990 499 499 150 150 10 10
(60.0%) (60.0%) (30.3%) (30.3%) (9.1%) (9.1%) (0.6%) (0.6%)
244 133244 1267133 198 490 298
(8.9%) (8.9%) (4.9%) (4.9%) (46.4%)
151 151 87 87 764 130 290 182
(9.2%) (9.2%) (5.3%) (5.3%) (46.3%) (7.9%) (17.6%) (11.0%)
(7.3%) (17.9%) (10.9%)
3/14
p Value p Value 0.8819 0.8819
Effects of Co-Worker and Supervisor Support on Job Stress and Presenteeism in an Aging Workforce: A Structural Equation Modelling Approach Tianan Yang 1 , Yu-Ming Shen 2 , Mingjing Zhu 3 , Yuanling Liu 4 , Jianwei Deng 5 , Qian Chen 6 and Lai-Chu See 7,8, * Received: 25 October 2015; Accepted: 18 December 2015; Published: 23 December 2015 Academic Editor: Cary Cooper 1 2
3 4 5 6 7 8
*
Department of Organization and Human Resource Management, School of Management and Economics, Beijing Institute of Technology, Beijing 100081, China; [email protected] Institute of Medical Informatics, Biometry and Epidemiology (IBE), Pettenkofer School of Public Health, Faculty of Medicine Ludwig-Maximilians University Munich, Munich 81377, Germany; [email protected] Institute of Psychology, Chinese Academy of Sciences, Beijing 100101, China; [email protected] Human Resources Department, Guangdong Women and Children Hospital, Guangzhou 511442, China; [email protected] Department of Public Administration, School of Management and Economics, Beijing Institute of Technology, Beijing 100081, China; [email protected] Medical Affair Department, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences, Beijing 100032, China; [email protected] Department of Public Health, Medical College, Chang Gung University, Taoyuan 33302, Taiwan Biostatistics Core Laboratory, Molecular Medicine Research Center, Chang Gung University, Taoyuan 33302, Taiwan Correspondence: [email protected]; Tel.: +886-3-211-8800 (ext. 5119); Fax: +886-3-211-8363
Abstract: We examined the effects of co-worker and supervisor support on job stress and presenteeism in an aging workforce. Structural equation modelling was used to evaluate data from the 2010 wave of the Health and Retirement Survey in the United States (n = 1649). The level of presenteeism was low and the level of job stress was moderate among aging US workers. SEM revealed that co-worker support and supervisor support were strongly correlated (β = 0.67; p < 0.001). Job stress had a significant direct positive effect on presenteeism (β = 0.30; p < 0.001). Co-worker support had a significant direct negative effect on job stress (β = ´0.10; p < 0.001) and presenteeism (β = ´0.11; p < 0.001). Supervisor support had a significant direct negative effect on job stress (β = ´0.40; p < 0.001) but not presenteeism. The findings suggest that presenteeism is reduced by increased respect and concern for employee stress at the workplace, by necessary support at work from colleagues and employers, and by the presence of comfortable interpersonal relationships among colleagues and between employers and employees. Keywords: co-worker support; equation modelling
supervisor support;
job stress;
presenteeism;
structural
1. Introduction Presenteeism can be viewed from two distinct perspectives. It is, of course, the literal antonym of absenteeism, and UK and European researchers in management, epidemiology, and occupational health [1–3] have investigated this positive aspect of the term, i.e., when employees remain at work, even when they are sick, or overstate their attendance, because of job insecurity due to downsizing
Int. J. Environ. Res. Public Health 2016, 13, 72; doi:10.3390/ijerph13010072
www.mdpi.com/journal/ijerph
Int. J. Environ. Res. Public Health 2016, 13, 72
2 of 15
and restructuring forces. However, presenteeism can be regarded as an indicator of lost work productivity, and US medical researchers and consultants [4] are concerned about the adverse effects that sickness and specific medical conditions might have on work productivity in organizations. In this article, we discuss presenteeism from the perspective of lost work productivity, because such analysis can address gray gaps between the absence of productivity and full productivity [5] and between healthy and unhealthy persons [6]. Moreover, previous studies suggest that further research is needed in order to better understand the potential differences between absenteeism and presenteeism [7–11]. Presenteeism was initially defined in the field of occupational health as the act of attending work while unable to perform effectively due to health problems [12,13]. The increasing attentions has been attracted is because presenteeism is responsible for 3 times and 1.8 times the financial burden of medical illness and absenteeism, respectively [14]. However, with the development of definition of presenteeism, more recently the definition has been extended to include other conditions and events that limit productivity, as suggested by Johns [15]. Advances in public health, medicine, science, and technology have greatly improved the health and life expectancy in developed and developing countries [16,17]. However, although retirement age is now older [18,19], physical and cognitive capabilities nevertheless deteriorate with age. This could lead to productivity loss and increased employer concerns regarding organizational competitiveness. Unfortunately, the extent of presenteeism among aging workers is not well understood. Employees who have demanding jobs, low decision latitude, job strain, and/or low social support tend to have more sickness absences [20] and, consequently, severe presenteeism [21–23]. Job stress occurs when an individual ability cannot meet job demands [24,25]. The job demand–resources model (JD-R) holds that when job demands are high and there are few job resources, employees may suffer more job stressors, which results in high job stress and other consequences. Because job demands must be satisfied if a worker is to remain employed, employees pretend to work hard at the workplace and forgo absences, even while they are sick or not fully productive. The greater the job requirements, the more efforts employees make in meeting them and the greater the probability that they will work while sick, to ensure full-time presence [26]. Fortunately, strong support from co-workers and supervisors improves work environments by relieving employee stress [27,28], which enhances job satisfaction and performance [29] and subsequently reduces presenteeism in enterprises and organizations [30,31]. Supervisors are in positions that can address employee complaints and help employees obtain necessary resources [32]. Co-workers can successfully finish work tasks and reduce stress and presenteeism [33]. In agreement with the buffering model of social support, Cummins [34] reported that employees that had good relationships with supervisors and co-workers are usually successful and productive at work, even when job stress is severe. Although co-worker support and supervisor support are both important in reducing job stress [35], most studies have investigated these two support mechanisms separately [36]. In addition, the relationship between co-worker and supervisor support has rarely been studied. Moreover, the level of co-worker and supervisor support that employees in an aging workforce receive is unclear. Most previous studies [27,34] used linear regression to investigate relationships between co-worker support, supervisor support, job stress, and presenteeism; however, such analysis cannot account for the complicated relationships among these variables. In this study, we used structural equation modelling (SEM) to examine the complicated effects of co-worker and supervisor support on job stress and presenteeism in an aging workforce. We hypothesized that there would be a direct buffer effect of supervisor support on job stress and presenteeism, a direct buffer effect of co-worker support on job stress and presenteeism, and a direct positive effect of job stress on presenteeism (Figure 1).
Int. J. Environ. Res. Public Health 2016, 13, 72 Int. J. Environ. Res. Public Health 2016, 13, 0072
3 of 15
Figure 1. 1. Initial Initial model model of of how how co-worker co-worker and and supervisor supervisor support support affect affect job job stress stress and and presenteeism. presenteeism. Figure
2. Method 2. Method 2.1. Data Data Source Source 2.1. We conducted conducted aa secondary secondary analysis analysis of of data data from from the the 2010 2010 wave wave of of the the Health Health and and Retirement Retirement We Survey (HRS) (HRS) in in the the United United States. States. The measures the the health, health, retirement, retirement, psychosocial psychosocial factors, factors, Survey The HRS HRS measures and productivity of the labour market and is funded by the National Institute of Aging and the Social and productivity of the labour market and is funded by the National Institute of Aging and the Social Security Administration of the United States. The HRS was initiated in 1992 and recruited persons Security Administration of the United States. The HRS was initiated in 1992 and recruited persons older than than 50 50 years years for for participation participation in in biennial biennial surveys surveys based based on on aa multistage multistage area area probability probability older sample. To avoid the aging problem and a decrease in the number of participants over time, new new sample. To avoid the aging problem and a decrease in the number of participants over time, samples were were added added every every66 years years [37,38]. [37,38]. In In 2006, 2006, the the HRS HRS added added aa participant participant lifestyle lifestyle questionnaire questionnaire samples (PLQ) to their core biennial survey, for a random 50% of the core panel participants. The The PLQPLQ was (PLQ) to their core biennial survey, for a random 50% of the core panel participants. developed by the HRS Psychosocial Working Group and includes a perceived ability to work scale was developed by the HRS Psychosocial Working Group and includes a perceived ability to work (PAWS), job stress scale, co-worker supportsupport scale, and supervisor support scale [39]. Detailed information scale (PAWS), job stress scale, co-worker scale, and supervisor support scale [39]. Detailed on the study population and survey methodology has been published elsewhere [40]. information on the study population and survey methodology has been published elsewhere [40]. In the the 2010 2010 wave wave of of the the HRS, HRS, PLQ PLQ data data were were available available for for 8080 8080 participants, participants, because because only only In respondents in in 2010 2010 who who had had completed completed the the face-to-face face-to-face PLQ PLQ interview interview in in 2006 2006 again again rotated rotated to to this this respondents mode of data collection [37,38]. Among these 8080 participants, 2730 (33.8%) were still employed and mode of data collection [37,38]. Among these 8080 participants, 2730 (33.8%) were still employed and older than than 50 50 years. years. Among Among those those still still employed, employed, 1649 1649 (60.4%) (60.4%) answered answered at at least least one one question question on on the the older “PLQ 2010”, and the percentage of those with missing data was less than 2.6%. Because presenteeism “PLQ 2010”, and the percentage of those with missing data was less than 2.6%. Because presenteeism is only only observed observed among among employed employed persons, persons, data data from from these these 1649 1649 participants participants were were analyzed analyzed in in this this is study (Table (Table1). 1). study Table 1.1. Demographic Demographic characteristics characteristics of of the the total total sample sample of of aging aging workers workers and and the the subset subset with with data data Table from the participant lifestyle questionnaire (this study) in the 2010 wave of the Health and Retirement from the participant lifestyle questionnaire (this study) in the 2010 wave of the Health and Retirement Survey,USA. USA. Survey, Characteristics Characteristics Sex Sex Male Male Female Female Age (years) 50–59 Age (years) 50–59 60–69 60–69 70–79 70–79 ≥80 ě80 Education NoEducation degree No degree Grade 1–9 Grade 1–9 High school diploma Two-year college degree Four-year college degree Master’s degree
Total Sample (n = 2730) Total Sample (n = 2730)
This Study (n = 1649) This Study (n = 1649)
1237 14931237 1493
(45.3%) (45.3%) (54.7%) (54.7%)
750 750 899 899
(45.5%) (45.5%) (54.5%) (54.5%)
1520 8641520 309864 37 309 37
(55.7%) (55.7%) (31.6%) (31.6%) (11.3%) (11.3%) (1.4%) (1.4%)
990 990 499 499 150 150 10 10
(60.0%) (60.0%) (30.3%) (30.3%) (9.1%) (9.1%) (0.6%) (0.6%)
244 133244 1267133 198 490 298
(8.9%) (8.9%) (4.9%) (4.9%) (46.4%)
151 151 87 87 764 130 290 182
(9.2%) (9.2%) (5.3%) (5.3%) (46.3%) (7.9%) (17.6%) (11.0%)
(7.3%) (17.9%) (10.9%)
3/14
p Value p Value 0.8819 0.8819
